[
https://issues.apache.org/jira/browse/HADOOP-14623?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16074965#comment-16074965
]
Allen Wittenauer commented on HADOOP-14623:
-------------------------------------------
I believe setting acks to 0 is intentional so that it doesn't block.
Additionally, metrics at scale are usually lossy (e.g., Ganglia using UDP) so
that the receiving end can actually handle the load.
> KafkaSink#init should set acks to 1,not 0
> -----------------------------------------
>
> Key: HADOOP-14623
> URL: https://issues.apache.org/jira/browse/HADOOP-14623
> Project: Hadoop Common
> Issue Type: Bug
> Components: common, tools
> Affects Versions: 3.0.0-alpha3
> Reporter: Hongyuan Li
> Assignee: Hongyuan Li
> Attachments: HADOOP-14623-001.patch
>
>
> {{KafkaSink}}#{{init}} should set ack to *1* to make sure the message has
> been written to the broker at least.
> current code list below:
> {code}
>
> props.put("request.required.acks", "0");
> {code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
